3D Face Landmark Conversion for Metaverse Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current metaverse systems fail to enable users to recognize and identify others when their avatars' 3D face models have changed, leading to difficulties in user recognition and interaction.
Innovation Solution
A method for generative converting 3D face landmarks, which accesses historical and current face models to blend facial characteristics, allowing the system to render a third face model that gradually changes from the original to the updated model, ensuring the user can be recognized by familiar individuals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If users change their avatar's 3D face model to pursue individuality and changing aesthetics, then the avatar's appearance adaptability is improved, but the user recognition reliability deteriorates
Solution Approach 1:
The system pre-extracts and stores facial characteristics from historical 3D face models before the user changes their avatar. These pre-extracted features serve as a reference baseline that enables recognition even when the face model changes completely, allowing the system to perform recognition without requiring the new face model to retain original visual characteristics.
Solution Approach 2:
The patent introduces facial characteristics extraction and comparison as an intermediary mechanism between the old and new face models. Instead of directly comparing visual appearances of potentially very different 3D models, the system extracts abstract facial features and uses these as intermediaries for comparison, enabling reliable recognition across significant appearance changes.
2Reliability
If the system stores and processes historical 3D face landmarks to maintain recognition, then the user identification capability is improved, but the system complexity increases
Solution Approach 1:
The system extracts only the essential facial characteristics from historical 3D face models and stores these extracted features separately from the complete face models. This extraction approach reduces the amount of data that needs to be processed and stored, simplifying the system while maintaining recognition capability.
Solution Approach 2:
Instead of storing and processing entire complex 3D face models for historical comparison, the system creates simplified copies in the form of extracted facial characteristics. These characteristic representations capture the essential recognition information while being much more compact and easier to process than full 3D models.
3Measurement precision
If the system generates blended 3D face landmarks by comparing historical and current models, then the recognition accuracy is improved, but the processing time increases
Solution Approach 1:
The system performs partial comparison by focusing only on key facial characteristics rather than analyzing every detail of the 3D face models. This selective approach to comparison maintains recognition accuracy by examining the most discriminative features while significantly reducing the computational burden and processing time.
Solution Approach 2:
The patent transforms the recognition problem from comparing complex spatial 3D coordinates to comparing extracted facial characteristic parameters. By changing the representation parameters from full geometric data to condensed feature vectors, the system achieves both high accuracy and efficient processing.
Data Source
AI summary
Embodiments of the present disclosure provide enhanced system and methods for implementing generative converting 3D face landmarks. An enhanced disclosed system and non-limiting method effectively renders a third 3D face model of a first user that enables a second user to easily recognize the first user, where the second user is only familiar with a first face model that is significantly changed in a second face model of the first user in a current interaction of the first user and second user. This method effectively renders the third 3D face model of the first user that can gradually change from the first face model to the second face model, and can be easily recognized by the second user.